TGF-beta inhibitor

TGF-Beta: A master immune regulator

Christopher Larson, Bryan Oronsky, Corey A. Carter, Arnold Oronsky, Susan
J. Knox, David Sher & Tony R Reid

To cite this article: Christopher Larson, Bryan Oronsky, Corey A. Carter, Arnold Oronsky, Susan J. Knox, David Sher & Tony R Reid (2020): TGF-Beta: A master immune regulator, Expert Opinion on Therapeutic Targets, DOI: 10.1080/14728222.2020.1744568
To link to this article: https://doi.org/10.1080/14728222.2020.1744568

Abstract

Introduction: Transforming Growth Factor-Beta (TGF-β) is a master regulator of numerous cellular functions including cellular immunity. In cancer, TGF-β can function as a tumor promoter via several mechanisms including immunosuppression. Since the immune checkpoint pathways are co-opted in cancer to induce T cell tolerance, this review posits that TGF-β is a master checkpoint in cancer, whose negative regulatory influence overrides and controls that of other immune checkpoints.

Areas Covered: This review examines therapeutic agents that target TGF-β and its signaling pathways for the treatment of cancer which may be classifiable as checkpoint inhibitors in the broadest sense. This concept is supported by the observations that 1) only a subset of patients benefit from current checkpoint inhibitor therapies, 2) the presence of TGF-β in the tumor microenvironment is associated with excluded or cold tumors, and resistance to checkpoint inhibitors, and 3) existing biomarkers such as PD-1, PD-L1, microsatellite instability and tumor mutational burden are inadequate to reliably and adequately identify immuno-responsive patients. By contrast, TGF-β overexpression is a widespread and profoundly negative molecular hallmark in multiple tumor types.

Expert Opinion: TGF-β status may serve as a biomarker to predict responsiveness and as a therapeutic target to increase the activity of immunotherapies.

Keywords: Checkpoint Inhibitors, Immune Regulation, Immunotherapy, TGF-β

Article Highlights
• TGF-β tumor promotion affects multiple processes including angiogenesis, fibrosis and immunosuppression via effects on the tumor microenvironment.
• TGF-β acts at multiple steps to suppress the generation of anti-tumor immune responses.
• Data is presented that supports the hypothesis that TGF-β is a master immune checkpoint, whose immunosuppressive properties have the potential to negate or attenuate the activities of checkpoint inhibitors.
• There are multiple inhibitors of TGF-β in preclinical and clinical development that target TGF-β and its signaling pathways that are promising “universal” checkpoint inhibitors for the treatment of cancer (e.g. small molecules, antibodies, antisense oligonucleotides, vaccines, oncolytic viruses).
• Given the favorable toxicity profile of these TGF-β inhibitors, combined with their ability to regulate checkpoint activity, they have the potential to synergistically enhance the efficacy of a variety of immunotherapies.
• In addition, TGF-β has the potential to serve as a predictive biomarker for response to cancer therapies.

1. Introduction

Transforming Growth Factor-Beta (TGF-β) is a diverse regulatory and fibrogenic protein with three isoforms, TGF-β1 (the most common), TGF-β2, and TGF-β3. TGF-β is secreted in a precursor form bound to a propeptide, cleaved by furin-type enzymes in the Golgi, transported to the extracellular matrix (ECM) in association with a latency associated peptide (LAP), and activated in the presence of diverse molecules such as thrombospondin-1, integrins, matrix metalloproteinases (MMPs), bone morphogenetic 1 (BMP-1) and reactive oxygen species (ROS) [1]. TGF-β impacts multiple processes including cell growth and differentiation, apoptosis, cell motility, extracellular matrix production, angiogenesis, and immune responsiveness[2–7]. These effects are very dependent upon context, including tumor hypoxia [8,9]. This pleiotropy results from activation of the TGF-β receptor, a heterodimeric complex composed of two transmembrane serine/threonine kinase receptors, TGF-βR1 and TGF-βR2 at the cell surface [10].

TGF-β binding leads to receptor transphosphorylation and activation, which, in turn, can phosphorylate the canonical receptor associated SMAD intracellular proteins that regulate target gene expression; e.g. (R-)SMADs, SMAD2/3), and/or initiate the less well understood non- SMAD signaling pathway through activation of mitogen-activated protein kinases (MAPKs), phosphatidylinositol 3-kinase (PI3K), TNF receptor-associated factor 4/6 (TRAF 4/6) and Rho- like GTPase intermediates [11]. The phosphorylated SMAD2/3 complexes with SMAD4 before translocation to the nucleus and initiation of SMAD-dependent gene transcription [7,12] (Figure 1). In addition, TGF-β1 signaling induces the expression of negative regulators, which are the inhibitory SMAD proteins, (I-SMADs) SMAD6 and SMAD7, that compete with R-SMADs for receptor interaction and mark the TGF-β receptors for ubiquitination and degradation [13].

As an inducer of cytostasis, protection and apoptosis, TGF-β initially acts to inhibit tumorigenesis [14], but later in the presence of oncogenic events and epigenetic perturbations can act as a tumor promotor [15].Perhaps the best example of this conversion of TGF-β into a cancer-promoting agent is the colon cancer Consensus Molecular Subtype (CMS) IV. Individual cases of colon cancer can be characterized by their gene expression profiles, and such analysis has revealed four common types of colon cancer (CMS I-IV), which correlate with clinical outcomes and response to therapy [16]. CMS IV is characterized by high levels of TGF-β and tumor fibrosis [17]. These tumors have a mesenchymal phenotype and are resistant to EGFR inhibitors, even in the absence of KRAS mutations. Patients with these tumors often present with advanced disease and have the worst 5 year survival of the four consensus molecular subtypes.

Unfortunately, patients with locally advanced but non-metastatic CMS IV cancers do not benefit from adjuvant therapy [18].The oncogenic events that lead to this switch in TGF-β‘s effects are varied. For example, mutations in SMAD4 and SMAD2 are relatively common in pancreatic, colorectal, endometrial, and gastroesophageal cancers, and block the growth inhibitory canonical signaling pathway, while leaving the non-canonical pathways intact, including growth-promoting MAPK and PI3K pathways [19].Activating mutations in the Ras pathway such as KRAS, NRAS, and HRAS are even more common [20] and lead to a similar effect by reducing Smad4 levels and blocking its ability to bind the TGF-β receptor complex and translocate to the nucleus [21]. Microsatellite instable colon cancers commonly lose TGF-β sensitivity completely due to frameshift mutations in the TGF-β receptor [22]. Regardless of the mechanism that cancer cells use to escape growth inhibition by TGF-β, tumor promotion is accompanied by angiogenesis, fibrosis, and immunosuppression by acting on the non-cancerous cells in the tumor microenvironment [23].

This review posits a role for TGF-β as a master immune checkpoint, whose profoundly immunosuppressive influence has the potential to negate or attenuate the activity of currently approved checkpoint inhibitors. Nearly all nucleated cells produce and respond to TGF-β, and of the myriad of mechanisms that cancer cells use to escape immune surveillance, TGF-β production has been described as the most potent [24,25]. On this basis, it may be appropriate to classify agents that target TGF-β and its signaling pathways for the treatment of cancer as universal checkpoint inhibitors.

2. TGF-β as a master immune regulator and checkpoint

Immune checkpoints, which include CTLA-4, PD-1, TIM-3, lymphocyte activation gene 3 protein (LAG-3) and B and T lymphocyte attenuator (BTLA), are expressed on activated T cells as built-in controls to countermand the co-stimulatory mechanisms shown in Figure 2 which might otherwise lead to the development of autoimmunity. One strategy for tumor cells to resist, avoid or suppress the immune response is to upregulate the expression of ligands that trigger these negative signals, such as PD-L1 and PD-L2 on tumor cells and the surrounding stroma.

The recognition of the centrality of immune checkpoints in the treatment of cancer has led to an immunotherapy paradigm shift in which checkpoint inhibitors are at the forefront.Checkpoint inhibitors that have received FDA approval include the anti-CTLA-monoclonal antibody (mAb) Ipilimumab, the anti-PD-1 mAbs, Nivolumab and Pembrolizumab and the anti- PD-L1 mAbs, Durvalumab and Avelumab. Despite the elicitation of long-term durable remissions in some patients with currently approved checkpoint inhibitors (CIs) in cancers such as melanoma, renal, lung, gastric, and hepatocellular carcinoma, the majority of patients do not benefit [26–28], which is problematic from the perspective of 1) toxicity since immune related adverse events are not only relatively common but also potentially life threatening [29–31] and 2) the potential for increased progressive disease since PD-1/L1 inhibitors have been linked with accelerated tumor growth [32,36].

One factor that is consistently associated with non-responsiveness to checkpoint inhibitors is an “immune excluded” tumor microenvironment, promoted by TGF-β signaling, in which T cells localize to the collagen- and fibroblast-rich tumoral periphery rather than to the epithelial tumor mass. In a melanoma study, which used large scale genomic and transcriptomic analysis to explain the differential clinical response patterns to anti-PD-1 therapy, the TGF-β pathway was identified as the key mediator of checkpoint inhibitor resistance [37]. Similarly, TGF-β signaling was correlated with lack of response to anti-PD-L1 (atezolizumab) therapy in patients with metastatic urothelial cancer, while preclinically the co-inhibition of PD-L1 and TGF-β converted immune-cold mouse tumors to immunologically-active ones, resulting in tumor regression, as shown in Figure 3 [38]. Also in T-cell excluded mouse models, immune checkpoint-resistant microsatellite stable (MSS) colorectal and liver tumors were rendered susceptible to anti-PD-1–PD-L1 therapy with TGF-β blockade [39]. On this basis, combinations of the TGF-β inhibitor, Galunisertib [40], with the PD-1/L1 checkpoint inhibitors, Durvalumab or Nivolumab, are under clinical investigation in metastatic pancreatic cancer (NCT02734160), NSCLC, (NCT02423343), hepatocellular carcinoma, and glioblastoma (NCT02423343).

TGF-β is generally described as a cytokine because unlike most checkpoints it is 1) secreted instead of membrane-bound, and 2) functions as a multifunctional regulator of diverse cellular processes [41], including, but not limited to, the inhibition of immune and CD8+ T-cell responses. However, in the manner of a “master” immune checkpoint, TGF-β blocks, or puts a “check” on multiple steps in the cancer-immunity cycle, as described by Chen and Mellman [6,7,42], which includes release of cancer cell antigens, cancer antigen presentation, priming and activation, trafficking of T cells to tumors, infiltration of T cells into tumors, recognition of cancer cells by T cells and killing of cancer cells. Due to this pleiotropy, as overviewed below (Figure 3), attenuation of the TGF-β master ‘OFF switch’ serves to normalize the abnormal tumor microenvironment and to ‘turn on’ the development of antitumor immunity. Hence, TGF- β inhibitors, which include antisense oligonucleotides, monoclonal antibodies like Fresolimumab, small molecules like Galunisertib, oncolytic viruses that encode a TGF-β trap, and bifunctional antibodies are potentially classifiable as checkpoint inhibitors (See Table 1) of the cancer immunity cycle as a whole. In addition to the bifunctional fusion protein targeting PD- L1 and TGF-βR-II (M7824) [43,44], another bifunctional protein targeting CTLA-4 and TGF- βR-II, has been developed, which could potentially inhibit the differentiation of Tregs and Th17 cells, and increase tumor-specific IGN-g+ effector and memory cells [45].

3. The Role of TGF-β in the Cancer-Immunity Cycle

Step 1: TGF-β Hinders the Release of Cancer Cell Antigens

Two prominent types of cell death are: apoptosis and necrosis [46,47]. The latter is inherently immunogenic and elicits inflammatory responses due to membrane rupture and the release of intracellular components such as cytokines and DAMPs (Damage-associated Molecular Patterns) [48–50], whereas the former is intrinsically tolerogenic, since macrophages phagocytose apoptotic cells and, hence, proinflammatory/proimmunogenic intracellular contents are not released. TGF- β induces apoptosis in many cell types such as lymphocytes and hepatocytes [51– 52] and, in addition, TGF-β is generated by apoptotic cells (ACs) and macrophages that engulf ACs [53,54], akin to an anti-inflammatory feed-forward loop, which hinders neoantigenic release. TGF-β blockade has been shown to reverse the suppressive effects of apoptotic cells on inflammation and adaptive immunity [6,55,56].

Step 2: TGF-β Subverts Dendritic Cell Function

Dendritic cells (DCs) are the main professional antigen-presenting cells (APCs) that shape immune responses either by activating naïve effector T cells through co-stimulation from molecules such as CD40, CD80 and CD86 or by establishing tolerance [57]. In the presence of TGF-β, DCs are induced to adopt a tolerogenic phenotype through downregulation of proinflammatory cytokines and cell-surface expression of their antigen-presentation machinery, including processed antigen peptide-loaded MHCII and co-stimulatory molecules, which interfere with antigen presentation [58,59] and result in cell death or anergy of the corresponding T cells. In addition, [56] tolerogenic DCs actively induce Foxp3, a transcription factor, which is required for the development of Tregs [60] as well as inhibitory molecules like PD-L1, CD95L, IDO, TGF-β and IL-10 [61].

Step 3: TGF-β Suppresses the Activation of CD8+ T-Cells

TGF-β inhibits T cell proliferation, activation, and effector functions, likely in part due to the downregulation of dendritic cell function [62] and partially due to suppression of IL-2 production, since the addition of exogenous IL-2 has been shown to partially reverse TGF-β- mediated suppression [63,64]. Regulatory T cells (Tregs) are strong inhibitors of T cell responsiveness that are induced by TGF-β signaling during T cell receptor activation [64,65], and their suppressive effects are blocked by neutralizing antibodies against TGF-β and IL-10 [66], with both playing a role in CD4+ CD25+ FoxP3 Treg induction [67]. Myeloid-derived suppressor cells (MDSCs) are another type of cancer-associated cell that has attracted interest as a target for immunotherapy due to its strongly suppressive activity, and TGF-β contributes to the expansion of MDSCs (particularly monocytoid MDSCs) [68] and mediates suppression of not only T cells but also natural killer cells [69,70]. TGF-β also induces the production of indoleamine 2,3-dioxygenase (IDO) [71,72], an enzyme that decreases tryptophan levels in the tumor, and thereby inhibits T-cell proliferation [73], and facilitates immune evasion [74].

Steps 4 and 5: TGF-β Reduces T Cell Trafficking and Infiltration to Tumors

T cell tumor-infiltration, a hallmark of so-called “hot” tumors, is generally correlated with a good prognosis. Conversely, a paucity of tumor T cell infiltration, a hallmark of so-called “cold tumors” is associated with therapeutic resistance and a poor prognosis [75]. Cancer associated fibroblasts (CAFs) are normal fibroblasts that take on pro-tumorigenic activity in response to stimuli within tumor microenvironment, with TGF-β playing a prominent role in inducing the CAF phenotype and mediating the CAFs’ activities, including driving cancer invasiveness and metastasis through epithelial-to-mesenchymal transition [76–79]. CAFs also affect T cell infiltration – high levels of CAFs are associated with decreased levels of infiltrating CD8 T cells and increased FoxP3 positive Tregs in humans [80], suggesting that CAFs play a causative role [78,81,82]. TGF-β, as a proangiogenic factor, upregulates VEGF and contributes to the formation of abnormal tumor vasculature [83], with resultant poor tumor perfusion and high interstitial pressure, which inhibits T cell infiltration, and drug delivery into tumors [84,85]. In addition to neoangiogenesis, TGF-β also drives a fibrotic response known as tumor desmoplasia, which is particularly present in solid tumors such as breast cancer and pancreatic ductal adenocarcinoma and serves as a barrier to T cell entry through compression of intratumoral blood vessels and elevation of interstitial pressure [86–88]. Ex vivo models have demonstrated that collagen degradation increases T cell infiltration [89]. Furthermore, collagen can influence tumor cell behavior through multiple mechanisms [90] and collagen can regulate the activity of tumor-infiltrating T cells [91].

Steps 6 and 7: TGF-β Antagonizes Recognition of Cancer Cells by T Cells and Hinders Killing of Cancer Cells

A hallmark of tumor development is T cell tolerance and impaired immune adaptive responses [92]. Tumors induce T cell tolerance through several factors, among them downregulation or complete loss of MHC class I molecules [93] from exposure to molecules such as TGF-β [94]. The mechanisms by which TGF-β suppresses CD8+cytotoxic T cell (CTL) effector functions are diverse and include inhibition of perforin, Granzyme B and A, interferon-gamma (IFN-γ), FAS ligand (FASL) expression, generation of T-regulatory cells, promotion of M2-macrophages (which secrete other immunosuppressive cytokines such as IL-10 [6,7,95,96], and epithelial-to- mesenchymal transition (EMT) induction [7,97]. TGF-β-induced EMT, in which cells lose their epithelial characteristics, including polarity, and acquire a stem cell like phenotype [98,99] is linked with downregulation of antigen presentation-associated pathway proteins and resistance to T-cell mediated killing. Accordingly, inhibition of TGF-β has been reported to have a variety of antitumor effects [95], an example of which is the systemic neutralization of TGF-β in vivo that was shown to result in tumor eradication in mice mediated by an increase in CD8+ T-cell mediated tumor-cell-specific cytotoxicity [100]. In humans, some cases of Hodgkin lymphoma is associated with Epstein-Barr virus (EBV) gene expression and is usually curable with conventional therapy. The role of TGF-β blockade in seven patients with EBV-positive Hodgkin lymphoma refractory to conventional chemotherapy was studied using autologous anti-EBV T cells that were modified to express a dominant-negative form of the TGF-β type II receptor to render them insensitive to TGF-β [101]. In this small trial, two patients had complete responses lasting over 5 years, and the remainder had partial responses or stable disease lasting 4, 6, 10, 13, and 19 months. Progressive disease was associated with loss of EBV antigen expression in the escaping Hodgkin lymphoma cells in one patient, and with lack of dominant-negative TGF-β receptor transgene expression in infiltrating lymphocytes in another.

4. TGF-β as a direct or indirect biomarker for treatment activity

The biomarker potential for TGF-β is complicated because protein-bound and non-protein bound, or free forms exist, and only free TGF-β ligand is available for binding to its receptor and, hence, biologically active. Nevertheless, high levels of TGF-β are associated with therapeutic resistance. In the phase 2 trial of Galunisertib in hepatocellular carcinoma, circulating TGF-β1 was measured. The median OS of patients with a >20% reduction of circulating TGF-β1 was 21.8 months vs. 7.91 months for patients with reduction of <20% (p=0.002) [102]. SMAD phosphorylation has also been used as a biomarker since TGF-β signaling is SMAD-mediated. In the phase I of Galunisertib, pSMAD2 reduction was observed in 9/14 (64%) of patients [103]. Alternatively, certain tumors are known to be associated with elevated TGF-β levels and the presence of desmoplasia, a surrogate for abnormal TGF-β pathway signaling and microenvironment remodeling and may serve as a predictive biomarker for the activity of TGF-β inhibitors (Figure 4). Desmoplasia, which refers to a dense extracellular matrix (ECM) of fibrillar collagen, hyaluronan, fibronectin, proteoglycans and tenascin C [104] compresses intratumoral blood vessels and leads to hypoperfusion, which, in turn, impedes the delivery of therapeutic agents, as well as rendering tumors hypoxic and treatment-resistant. TGF-β promotes nutrient uptake and extracellular matrix production. The tumor types that are associated with a high degree of fibrosis or desmoplasia, to which TGF-β is a key contributor, include pancreatic ductal adenocarcinoma (PDAC), the CMS4 molecular subtype of colorectal cancer [17], breast cancer, hepatocellular carcinoma (HCC), and various sarcomas[105], which therefore may benefit preferentially from TGF-β inhibition, especially since radiotherapy and chemotherapy induce TGF-β activity. The antihypertensive angiotensin receptor blocker, losartan, is associated with reduced TGF-β expression, and retrospective analysis of ARB-treated PDAC patients demonstrated survival of ~6 months longer than non-ARB-treated patients [106]. This data was the basis for an ongoing phase II clinical trial with losartan and FOLFIRINOX in PDAC (ClinicalTrials.gov identifier NCT01821729). Additionally, through stromal modulation of known desmoplastic tumors, the TGF-β neutralizing antibody, 1D11, improved the distribution and efficacy of DOXIL in a breast carcinoma mouse model [85]. In a mouse pancreatic model, pirfenidone, a TGF-β inhibitor clinically approved for the treatment of idiopathic pulmonary fibrosis, more effectively suppressed tumor growth in combination with gemcitabine than gemcitabine alone, with evidene of suppression of desmoplasia [88,107]. 5. Toxicity of TGF-β Inhibitors In animals, small molecule inhibitors of TGF-β, like Galunisertib, are associated with valvulopathies and aneurysms of the ascending aorta [108], toxicities which mimic the connective tissue disorder Loeys–Dietz Syndrome (LDS), a subset of Marfan Syndrome, characterized by a disruption of TGF-β signaling with structural changes of heart valves, aortic dilatation, dissection and rupture [109]. Fortunately, however, in the first in human study of Galunisertib, no medically significant cardiac toxicities were observed. The main drug related adverse event with Fresolimumab (formerly GC1008), a pan-TGF-β ligand monoclonal antibody that has been tested in Phase 2 trials, is reversible non-malignant keratoacanthomas [110] not cardiovascular toxicities. Cardiac safety and acceptable safety in general have been demonstrated with Trabedersen [AP12009], an antisense oligonucleotide targeting TGF-β2 and Belagenpumatucel-L, a TGF-β2 antisense gene-modified allogeneic cancer cell vaccine [111,112] . In the trial of autologous T cells modified with a dominant-negative TGF-β receptor, neither cardiotoxicity nor cytokine release syndrome (a common side effect with CAR-T cells), or chronic inflammatory disease (seen in a mouse model with the dominant negative T cell receptor) [113] was encountered, with absence of the latter despite persistence of the T cells for up to 51 months, possibly attributable to the fact that the investigators used only EBV-specific T cells in this particular disease. The potential (or lack of potential) for cardiac toxicities with TGF-β inhibitors is especially important in light of planned combinatorial strategies with PD-1/PD-L1 inhibitors, since several cardiotoxic events such as myocarditis, HF, heart block, heart failure, myocardial fibrosis and cardiomyopathy [114,115]have been recently been recognized with these checkpoint inhibitors [95]. The incidence of cardiotoxicities with checkpoint inhibitors may even be underestimated since troponin levels, for example, are not routinely assessed in oncology trials. 6. Conclusion Strategies to transform immunologically non-inflamed or “cold” tumor microenvironments into inflamed or “hot” ones through the use of checkpoint inhibitors (CIs) in combination with other checkpoint inhibitors, molecularly targeted therapies, vaccines and CAR- T cells have to date met with limited success, benefiting only a minority of patients [116] and often at the expense of high and potentially life threatening toxicities [117]. The advantage of using TGF-β inhibitors in combination with checkpoint inhibitors is their low apparent toxicity combined with their ability to negatively regulate immune checkpoint activity for potential synergistic anti-cancer benefit, as shown below in Figure 5. A number of preclinical studies have demonstrated an additive or synergistic anti-tumor effect by combining TGF-β blockade with checkpoint inhibitors [95,38]. This dual blockade approach is now being tested in clinical trials in a variety of tumor types [2]. The disadvantage is that TGF-β signaling also elicits protective or tumor suppressive effects, at least in early stage cancer [118] and, therefore, TGF-β inhibitors also in theory have the potential to accelerate tumor progression.TGF-β has been referred to as a “jack of all trades” [119] at least with regard to the immune system and cancer because of its multifaceted role and control over T cell lineage commitment, angiogenesis, immune suppression, and tolerance, which suggests that it is really more of a master checkpoint or “jack of all checkpoints” whose inhibition has the potential to render tumors more immunogenic and, hence, facilitate both innate and adaptive immune activation. This definition of TGF-β as a master immune regulator is more than simply academic insofar as the current trend in immunotherapy is to combine agents that stimulate and expand immune cells, a strategy that is likely of limited value if T cell exhaustion, an immunosuppressive tumor microenvironment (TME) and “cold” tumors with a paucity of cytotoxic lymphocytic infiltrate (all hallmarks or signatures of abnormal TGF-β signaling) are simultaneously present but not addressed with inhibitors of immunosuppressive molecules such as TGF-β. Although an area of active investigation, there are no validated, reliable predictive biomarkers to determine whether patients are likely to respond to checkpoint inhibitors, since changes in immunological markers such as immune cell subsets, tumor mutational burden, PD- 1/PD-L1 expression, microsatellite status and DNA mismatch repair do not always correlate with patient outcomes or response to CIs [120]. Since abnormal TGF-β signaling and immunosuppression are such universal features of advanced cancers, serial measurements of microenvironmental factors indicative of their presence such as intensity and phenotype of immune infiltrate, angiogenesis, sclerosis/desmoplasia, SMAD phosphorylation and mesenchymal molecular features have the potential to help determine which patients are likely to respond to immunotherapy and which are not, and merit further study. Ultimately, it remains to be seen whether TGF- β inhibitors, such as antibodies, small molecules, antisense RNA, vaccines and oncolytic viruses that overexpress TGF-β binding ligands during replication, are able to convert cold tumors into hot tumors more responsive to immunotherapy ,and, whether, as this review posits, TGF-β is a master immune regulator, that plays a dominant role in the susceptibility of tumors to immune checkpoint inhibition. 7. Expert Opinion TGF-beta is a master immune regulator and “immunosuppressive brake” situated at the apex of T-cell and non-T-cell associated immune checkpoints including LAG-3, TIM-3, TIGIT, VISTA, BTLA, B7-H3, KIR, PI3K, and CD47, which collectively drive a growth-promoting tumor microenvironment and escape from immune surveillance [6,7,95]. On the one hand, tumors are complex multicellular systems, and, therefore, it is perhaps overly simplistic and reductionist to elevate TGF-beta above other immune checkpoints in terms of its “command and control” function. On the other hand, regulatory networks in biology are intrinsically hierarchical [121] and tumors themselves are hypothesized to be hierarchically organized, similar to the tissues from which they arise [122]. Therefore, TGF-beta, which broadly induces immune tolerance and counteracts the immunostimulatory effects of checkpoint inhibitors, is inferred to broadly serve as an OFF switch. It not only mutes or terminates cytotoxic innate and adaptive responses on its own, but also exercises overriding control over a wide swath of cancer hallmarks including epithelial-mesenchymal transition (EMT), angiogenesis, and stromal fibrosis, far more so than the widely characterized PD-1/CTLA-4 axis, which functions in the capacity of a single master immune checkpoint. As currently approved, checkpoint inhibitor monotherapy is associated with low response rates. In combination (e.g. anti-CTLA-4 and anti-PD-L1) higher response rates are observed, but at the cost of significantly more toxicity [123]. The result is that for all hype and hope [124] with which checkpoint inhibitors are associated, and despite unprecedented and clinically relevant anticancer activity in a subset of tumor types, most patients do not respond, and in the majority of patients that do respond the benefits are not durable. Currently, no reliable and validated biomarkers are available to predict which patients would be most likely to benefit from these expensive therapies that are characterized by relatively high or at least non-negligible rates of grade 3, grade 4 and even grade 5 immune toxicities. Therefore, there is a substantial unmet need to development new treatments and new strategies, which in combination with immune checkpoint blockade and/or other modalities, improve the magnitude and the duration of the anti- tumor immune response, and prevent or overcome resistance. As a corollary to the above, clinically meaningful biomarkers are needed for the identification of patients that are most likely to benefit from treatment and to monitor therapeutic responses. Promising data from preclinical studies and early clinical trials of the combination of TGF-beta inhibitors with checkpoint inhibitors (and other therapies such as radiation therapy), suggest the potential for this therapeutic strategy to “normalize” the immunosuppressive tumor microenvironment for more effective PD-1/PD-L1 inhibition [2,3,38,125–127], as well for sensitization to other immunotherapies and conventional treatments. Immune enhancement with checkpoint inhibitors in the absence of immune normalization i.e., conversion of immunosuppressed tumors to non-immunosuppressed ones, has a limited ratio of benefit to risk. Hence, the basic biological rationale to combine TGF-beta inhibitors with PD-1/CTLA-4 inhibitors is three-fold: 1) the former, as a negative regulator of checkpoint activity, is thought to gate or control the responsiveness to the latter 2) the toxicity and mechanisms of action of both therapies are non-overlapping and 3) serum and tumor expression of TGF-beta levels may predict the likelihood of response to and potential benefit from treatment. It is anticipated that in the next 5 years, based on the tolerability and broad potential of TGF-beta inhibitors to sensitize checkpoint inhibitors, a large number of combination trials will be initiated, in an expanding number of different tumor types, perhaps also in combination with other modalities such as radiation therapy, whose pro-immunogenic effects may result in a more effective in situ tumor vaccine with enhanced memory T-cell responses. These combination therapies will be studied not only in advanced/metastatic disease, but also in the adjuvant setting, with a focus on optimization, based on an in-depth understanding of the underlying biology. One potentially noteworthy TGF-beta inhibitor, called AIM-001, which has received FDA approval to begin a Phase 1 trial in combination with a checkpoint inhibitor, is a tumor-specific replicating oncolytic adenovirus that overexpresses a TGF-beta trap transgene, which binds to and prevents TGF-β1 from binding to TGF-β receptor ΙΙ. In patients with advanced tumors that received these same oncolytic adenoviruses under compassionate use protocols dramatic and durable anticancer activity was observed along with T-cell infiltration of tumors, indicative of conversion to an inflamed tumor phenotype.

Hopefully, within 5 years, one or more of these TGF-beta-immunotherapy combinations, will be integrated or incorporated as standard of care potentially in traditionally “cold” or non-T- cell-inflamed tumors such as colorectal, ovarian, prostate and pancreatic ductal adenocarcinoma. It is also predicted that TGF-beta itself will serve as a validated predictive biomarker either alone or in combination with other biomarkers, that will facilitate the selection of patients most likely to benefit from the impairment of the tumor-supportive and immunosuppressive effects of TGF- beta.

Funding

This paper was not funded.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.